基于Scikit-Learn的垃圾短信过滤方法实证研究

An Empirical Study on Spam Messages Detection Method Based on Scikit-Learn

文章为有效应对垃圾短信,在短信数据集"SMS Spam Collection"上,以Scikit-Learn为工具,通过实验对比验证,结果表明,在比较的7种垃圾短信过滤统计学习方法中,朴素贝叶斯和支持向量机方法在判别准确率方面明显优于其他方法,这2种方法可以作为其他方法用以比较的基准测试方法.

In order to effectively deal with spam messages,this paper is verified by experiment in the SMS da-taset'SMS spam collection'. It is found that in the experimental comparison of seven kinds of spam filteringstatistical method,naive Bayesian and support vector machine in the discriminant accuracy was significantlybetter than the other methods.These two methods can be used as a benchmark test(baseline)method for theother methods.